The broad, long-term objectives of this proposal are to quantitatively characterize, compare, and simulate the use of task space control in normal and pathological gait due to lower limb amputation. Extensive research has detailed joint kinematics and kinetics of human locomotion and movement control with normal and various motor function disorders. Yet, relatively little is known about how humans relate joint-level information to higher-level locomotor goals. Understanding basic biomechanical tenets governing leg control before and after injury will provide a general theoretical framework for creating new therapeutic interventions for pathological gait disabilities. This project proposes to use behavioral locomotion experiments to study task space control in both healthy intact human subjects and in both Below-Knee (transtibial) and Above-Knee (transfemoral) amputees walking on prosthetic legs. This project will use extensive data analysis to quantify and characterize task space control during locomotion using variables identified from biomechanical models and by eigenanalysis. Based on the analytical results, a further study will create a physics-based computer simulation of pathological amputee gait with corrective joint torques computed from the differences discovered between normal and amputee gait. The general hypothesis of the project is that humans control locomotion by precisely guiding their legs through a predefined task space while allowing flexibility in joint-level dynamics and that more proximal limb amputations will result in greater losses in this ability to exploit task space redundancy. Three specific aims to study the use of task space control during locomotion are addressed: normal human locomotion (Aim1); pathological gait due to lower limb amputation (Aim2); and, the evaluation and simulation of the changes from normal to pathological human walking (Aim3). Achieving these three aims will provide immediate deliverables by quantifying functionally relevant task variables for normal human gait and a theoretical understanding of how these control variables change after a neuromechanical pathology such as lower limb amputation. This work will develop new analytical tools to quantitatively assay normal and amputee gait, however, it will also be generally applicable to the study of all gait pathologies. It could further impact future therapeutic approaches for gait rehabilitation that use task-level rather than joint-level rehabilitation goals and inspire control systems for the rapidly growing field of rehabilitative robotic devices. The proposed work will also form a theoretical framework for future studies of the neural mechanisms underlying leg control during locomotion.